@Dashrender said in Win10: Parked CPU -x:

I have never manually managed this.

Nor have I.

@dafyre said in Collectl Performance Monitoring for Linux:

@scottalanmiller said in Collectl Performance Monitoring for Linux:

@dafyre said in Collectl Performance Monitoring for Linux:

@JaredBusch said in Collectl Performance Monitoring for Linux:

@scottalanmiller said in Collectl Performance Monitoring for Linux:

Oh I didn't catch that. Yes, one or the other. Both would be a LOT of traffic and overhead and who would use both?

@dafyre obviously.

What do you do if you need to visualize all this data that collectl is gathering? I see no utilities included to help with that... unless it is used for on the fly tracking, and not long term stuff.

ELK can collect directly, it doesn't need Collectl. At least for most things.

How? Using the Logstash fowarder?

Topstash or something like that.

@Dashrender said:

@scottalanmiller said:

@Dashrender said:

@scottalanmiller said:

@Dashrender said:

Scott, you mentioned that this is all at the kernel level - could you roll your own version of this?

Sure, you'd have to write your own storage layer, though. So it's not trivial in any way.

OH.. that's where I was confused I guess... I thought the storage layer was part of KVM (that's the hypervisor they use, right?)

Not part of KVM itself. The Scale HC3 is unique, there is no software version available on the market.

So they wrote the storage layer? Cool - good to know/understand that.

Yes, Scale is primarily a storage vendor. Before they made their Hyperconverged product, they made scale out storage only. That was before KVM was mature enough to make the HC3 product. They no longer sell the storage layer, it is now developed purely and designed solely around the needs of the HC3 product so is completely unique to that. It's the storage layer and the storage integration (and support) that are their selling points. That's what makes them special and unique. KVM and the hardware on its own you could do yourself and you could easily make due with a different interface.

@DustinB3403 said:

Reading the below stats is a bit, awkward to say the least it shows the information, but it's not entirely clear in my opinion. XenCenter seems to graph the disk usage more accurately at the Hypervisor level. (shown below is XO showing the Hypervisor level)

0_1452189348096_2016-01-07_12-55-36.png

When you select a VM though you get something similar to below.

0_1452189500086_2016-01-07_12-58-12.png

Now I personally, don't care so much about an individual VM's disk performance, I care about the host. So for me using XenCenter provides a "better" explanation of what I'm reading.

That's pretty cool. However keep in mind that in some cases, a single VM's disk performance can be hurting performance for the other VMs as well.